All field observations of live aboveground biomass in tropical (and nontropical) forests are indirect, un-validated estimates for just the larger stems (EAB, estimated aboveground biomass). For multiple reasons (see below), it remains unclear how the existing EAB values for this biome can best serve the models.

To derive EAB, all live stems in a stand above some diameter limit (usually 10 cm) are measured for diameter (rarely also height). Each stem's aboveground biomass is then estimated using an allometric relationship between biomass and diameter (sometimes also height) that was derived by harvesting and weighing individual trees at another site(s). This approach raises the issue of “…misplaced concreteness” with respect to forest biomass estimates (Clark and Kellner, 2012). Different allometric equations can produce starkly different values of EAB from the same set of stem measurements; this is illustrated in Table 3 by the range of the five estimates (242–428 Mgha-1) produced by different allometries but from the same 1992 set of tree diameter inventory data at the NOU-PP site. To determine which, if any, of such estimates is accurate for a given landscape would have required structured follow-up harvests at the site to test the applicability of a given allometric relation to that forest (Clark and Kellner, 2012). Because as yet no such validation has been carried out in a tropical forest, all EAB values for this biome are highly uncertain at the site level. While the range of these estimates is the only available guidance for upper and lower bounds for this biome, the accuracy of this range is also unknowable. Given these uncertainties, it will be important to maintain the actual field data (e.g., diameter and taxonomy of all stems) in a publically accessible archive, so that users could apply alternative allometries or estimation methods in the future.

For testing models against field observations of tropical-forest biomass (see Cleveland et al., 2015), a separate important issue is the within-forest spatial heterogeneity of EAB. For example, within a 10 ha area of French Guianan forest where EAB averaged 301 Mgha-1 (NOU-GP in Table 3) the range of the estimates for individual hectares was 230–416 Mgha-1 (Chave et al., 2001). A similarly large range among individual hectares was also found within the 50 ha plot on Barro Colorado Island, Panama (180–440 Mgha-1; Chave et al., 2003). Due to this local-scale variation, landscape-scale biomass observations would be required for most types of model–data fusion (except in the case of individual-based and forest demographic models (e.g., Hurtt et al., 2004), which explicitly incorporate this spatial heterogeneity).

Many models, particularly those that simulate forest demographics, use allometric equations to relate stem diameter to biomass. They also typically use estimated production of woody biomass to calculate diameter increments. In such cases, comparisons of both biomass and diameter increment for the same forest are therefore only sensible if the same allometric scaling is used. Again, detailed knowledge both of the data products (including EAB) and of model structures is critical.

Current ILAMB benchmarks for tropical regions include maps of aboveground biomass across the biome based on remote sensing products (e.g., Saatchi et al., 2011; Baccini et al., 2012). Large divergences between these maps (Mitchard et al., 2014) highlight the unresolved uncertainties due to method issues for both the remotely sensed data and the field observations (e.g., un-validated allometries, landscape-scale samples vs. a single 1 ha plot).

Estimates of tropical-forest CWD span a wide range and are method-dependent (see Table 4). The different methods in current use can produce significantly different estimates for the same site and time (e.g., the two 2005 estimates for JH-CLAY, Table 4). The spatial heterogeneity of standing and fallen CWD within tropical forests calls for landscape-scale sampling. CWD stocks are also likely to significantly change through time due to the temporal variation in tree mortality in this biome (see below).

Highly replicated, landscape-scale field observations of this C stock are potentially useful as a lower bound. Fine-root biomass is notoriously heterogeneous on multiple spatial scales. Studies within diverse tropical forests have demonstrated within-forest decreases in fine-root biomass with increasing microsite-scale availability of nutrients or water, as occurs along catenas or among the intercalated soil types in these forests (Palmiotto et al., 2004; Powers et al., 2005; Epron et al., 2006; Espeleta and Clark, 2007; Kochsiek et al., 2013; Noguchi et al., 2014; Wurzburger and Wright, 2015). Also, landscape-scale fine-root stocks can vary markedly through time. For example, fine-root stocks varied by 2.5 Mgha-1 over a 7-year period in a Costa Rican wet forest (LS in Table 5; Espeleta and Clark, 2007). Dynamic ecosystem models would ideally hope to capture such time series.

As illustrated in Table 5, the methods used to quantify fine roots vary in multiple ways, including the maximum diameter of evaluated roots, the depth of soil cores, and whether or not dead roots are included. These method variations make cross-site comparisons and model benchmarking difficult.

A separate critical issue affects observations of fine-root stocks in all forest types, boreal to tropical: fine-root sampling in forests is usually restricted to the surface soils. No study has quantified fine roots all the way down the soil column in any tropical forest (see Table 5). The soils underlying these forests are often many meters deep. Nepstad et al. (1994) found live roots down to at least ca. 18 m depth under one Brazilian tropical forest (TAP-DROU in Table 5); over the depth interval 2–6 m, fine-root density was relatively constant but much reduced compared to that of surface fine roots. Given the great soil volume at depth, the contribution of deep fine roots both to total fine-root stocks and for ecosystem function may be significant in tropical forests. Models increasingly predict root stocks at different levels in the soil based on an assumed exponential decay down the vertical profile. In such cases model–data comparisons should be made for the actual soil layer of the measurements. Because all models require total root mass, however, extrapolation will be required in one domain or the other.

There are as yet no stand-level observations of coarse roots in any forest type. In tropical forests, the field sampling for these spatially variable organs has been confined to harvesting the root systems of selected individual trees (e.g., Niiyama et al., 2010) or to sampling coarse roots in pits or trenches away from trees, thus missing their tap roots and other large roots (e.g., Castellanos et al., 1991; Veldkamp et al., 2003). A recent survey of the available harvest data (Waring and Powers, 2017) found that root : shoot ratios for individual trees from old-growth tropical forests averaged ca. 0.65, indicating the importance of this biomass component. Notably, this ratio strongly contrasts with the 0.21 multiplier commonly used to extrapolate tropical-forest coarse-root biomass from estimated aboveground live biomass (e.g., Malhi et al., 2009; Girardin et al., 2010; Quinto-Mosquera and Moreno, 2017).

SOC is strongly underestimated in all forest types (boreal to tropical) because it is rarely if ever quantified to depth (Jobbagy and Jackson, 2000). The limited tropical data in hand for subsurface SOC indicate that total SOC can dominate the C inventory in lowland tropical forests, where soils are commonly several to many meters deep (Sombroek et al., 2000). In two tropical forests where SOC was quantified to at least 3–4 m depth (Table 6), the cumulative SOC stock to the maximum sampled depth was roughly 10 times that at the surface (0–10 cm). Notably, cumulative SOC also exceeded the estimated C in aboveground live biomass (Table 6). Only in one of these cases (LS-younger Oxisol) was SOC quantified down to the parent material. In the other two, the sampling ended many meters shy of the total soil depth, thus missing large amounts of SOC. At the Amazonian site PARAGOM, where Trumbore et al. (1995) sampled SOC down to 8 m (Table 6), the soil shafts of Nepstad et al. (1994) actually extended down to 18 m depth.

The incompletely quantified SOC is a particularly critical data gap for tropical forests. There is accumulating evidence that the huge C stocks in the deep soils underlying many of these forests are not inert (e.g., Trumbore et al., 1995; Veldkamp et al., 2003). At the Costa Rican LS site (Table 6), the SOC at 2–3 m depth was found to be strongly temperature-responsive (Schwendenmann and Veldkamp, 2006), indicating a vulnerability of this large tropical-forest C stock to future warming. Deep SOC (1–4 m depth) at this forest site was also found to mobilize with forest-to-pasture conversion (e.g., 30 Mg C ha-1 lost from this subsurface soil layer in ca. 30 years; Veldkamp et al., 2003). Changes in tropical-forest SOC, particularly in the deeper soil layers, could strongly impact the total forest C stocks and net C balance of this biome.

A second issue in tropical forests is that SOC shows marked spatial variation on all scales: from one square meter to the next (Powers, 2006) and across the major edaphic changes (topography, soil types; see Richter and Babbar, 1991) within a forest. An example of this within-forest heterogeneity is the significant difference in cumulative SOC content between two major soil types at the LS site (Table 6). Distributed and replicated sampling is therefore required to quantify this important C stock.

The eddy flux method has been criticized for uncertainty in its nighttime measurements. This is especially obvious in tropical areas, where nighttime turbulence is not well developed. Nevertheless, … Convincing results can be obtained from daytime eddy flux measurements… (Tan et al., 2013)

For NEE at longer time steps (days to years), however, estimates based on the eddy flux technique in tropical forests do not provide reference-level field benchmarks for the models. Multiple issues for this technique in these forests create large uncertainties about the magnitude and even the sign of such estimates. The prevalence of still air conditions at night (e.g., 70–80 % of 30 min nighttime periods; Loescher et al., 2003: Costa Rica; Miller et al., 2004: Brazilian Amazon) means that the technique is inoperative or likely to be strongly biased during most nighttime periods. Studies have shown that the terrain irregularities typical of tropical forests can produce artifacts due to CO2 movement into or out of an eddy flux site through lateral advection in these still air periods (Goulden et al., 2006; de Araújo et al., 2008; Tóta et al., 2008). In multiple studies (Araújo et al., 2002; Saleska et al., 2003; Miller et al., 2004) the eddy flux estimate of yearly NEE from a given year's worth of data switched from C source to C sink with different data filtering for these periods of slow air movement. Further uncertainty in eddy flux estimates of tropical-forest annual NEE is caused by the substantial data gaps due to heavy rainfalls, to frequent problems with instruments and with power, and to equipment damage from animals, treefalls, and lightning. For one forest eddy flux study in Borneo, the actual NEE data after data-filtering covered only 30 % of the 17-month study period (Katayama et al., 2013). Diverse methods are then used to fill the many periods of missing data (e.g., predicting daytime NEE based on radiation data, Katayama et al., 2013, or assuming a constant value for nighttime NEE, Loescher et al., 2003).

… there is no way of directly measuring the photosynthesis or daytime respiration of a whole ecosystem of interacting organisms; instead, these fluxes are generally inferred from measurements of net ecosystem–atmosphere CO2 exchange (NEE), in a way that is based on assumed ecosystem-scale responses to the environment. … Our [13C/12C] analysis indicates that daytime ecosystem respiration differed fundamentally from standard predictions that were based on nighttime NEE and temperature… (Wehr et al., 2016)

Although GPP estimates have been produced by tropical-forest eddy covariance studies, the sole CO2 flux that is actually assessed with this technique is NEE, the small difference between two much larger, opposing fluxes (GPP and Reco). As discussed above, eddy flux NEE data from tropical-forests are themselves highly uncertain and incomplete. The standard current approach for partitioning NEE into GPP and Reco is based on assumptions about forest ecophysiology that have recently been challenged by findings from parallel 13C / 12C measurements in a temperate forest (Wehr et al., 2016).

Alternatively, bottom-up biometric approaches have been used to estimate GPP for some tropical-forest sites (e.g., Doughty et al., 2014; Malhi et al., 2015). These studies, carried out in a single 1 ha plot per forest, have been based on combining sparse direct observations of some components of production and respiration with intuitive estimates for, or omission of, many unmeasured components (see Sect. 2 and Table 7). In tropical forests, the summed C in the unmeasured processes may equal a significant fraction of total GPP (Clark et al., 2001a; Litton and Giardina, 2008).

Similarly, existing eddy flux estimates for whole-forest respiration in this biome remain questionable due to multiple issues: (1) the uncertainty of the NEE estimate from which Reco is inferred (see above), (2) the likelihood of lost (and/or extra) respiration due to lateral advection of CO2 during the predominantly still nights (Goulden et al., 2006; Tóta et al., 2008), and (3) unresolved questions about the assumptions underlying the estimation of daytime Reco from NEE (Chambers et al., 2004; Wehr et al., 2016; Wohlfart and Galvagno, 2017).

Benchmark-level field observations of these two fractions of Reco are as yet lacking for tropical forests. Neither of these fluxes can be directly assessed in the field at the ecosystem level. Some estimates of stand-level Ra (e.g., Doughty et al., 2015, and included references) have been derived for different tropical forests in the Global Ecosystem Monitoring (GEM) project. These estimates were based on sparse field measurements in a single hectare of the studied forest of a subset of Ra components (fine-root respiration (estimated as soil CO2 efflux minus that with root exclusion), canopy-leaf dark respiration, and tree-bole CO2 efflux). These measurements were then combined with intuitive estimates for two unmeasured Ra components (daytime leaf respiration, respiration by coarse roots). The substantial CO2 efflux from small-diameter wood (<10 cm diameter) was not considered; however, in a Costa Rican forest this Ra component was estimated to account for 70 % of total woody CO2 efflux, based on extensive sampling from mobile climb-up towers (Cavaleri et al., 2006). In the soil, the intimate interrelations among roots, root exudates, root symbionts, and soil microbes make the distinction between Rh and Ra both conceptually and methodologically challenging (Trumbore, 2006). An aspect of Rh that is rarely measured in tropical forests is the CO2 efflux from decomposing coarse woody debris. This respiration component has been estimated at 6–16 % of total tropical-forest Reco, based either on extrapolating spot field measurements of respiration from CWD to the stand level (Chambers et al., 2004: Central Brazilian Amazon) or on combining landscape-scale estimates of CWD stocks with inferred CWD turnover time (Hutyra et al., 2008: Eastern Brazilian Amazon; Cavaleri et al., 2008: Costa Rica).

No benchmark field observations are available for total NPP. As is the case in all other forest types (Clark et al., 2001a), the field studies in tropical forests have been restricted to a subset of NPP components (Table 7). Those that remain unquantified could sum to a substantial fraction of total NPP (see also Clark et al., 2001a, b; Litton and Giardina, 2008; Cleveland et al., 2015). For the models, the sum of the NPP components assessed in the field provides a lower bound for total NPP.

Two NPP constituents missing from the field studies (Litton and Giardina, 2008) and from most models (Fatichi et al., 2014) so far are the amounts of new fixed C being lost (exported) from the plants belowground, either to root symbionts (nodules and/or mycorrhizae) or to the soil through root exudation. Isotopic evidence from a CO2 enrichment study in a temperate forest indicated the likelihood of significant C export from the roots; belowground transfer of a substantial fraction of the assimilated C was found, with strong signals in mycorrhizal sporocarps and in soil respiration (a mix of Rh and Ra) but not in the fine roots (Steinmann et al., 2004). Most tropical trees support mycorrhizae (Janos, 1980), and legumes, potential N-fixers, are present in most tropical forests. The possibility therefore exists of considerable allocation of NPP to symbionts. This aspect of C cycling is practically unstudied in the biome. In one exceptional study in a Costa Rican forest (Lovelock et al., 2004), extra-radical hyphal production by arbuscular mycorrhizae at 0–10 cm soil depth was estimated at 1.5–1.9 Mgha-1yr-1. Because the total plant-assimilated C going into new mycorrhizal fungal tissues also includes that incorporated into spores and sporocarps, the hyphae inside roots, and all the hyphae in the soil below 10 cm depth, this NPP component appears to be significant in this forest. Root exudation, as yet unstudied, is another potentially nontrivial portion of tropical-forest NPP. Another NPP constituent omitted from field C cycle studies is the production of volatile organic compounds (VOCs). Guenther et al. (1995) found total annual VOC emissions from tropical forests (isoprene, monoterpenes, other reactive VOCs, and other VOCs combined) to reach 75 g C m-2, but with uncertainties greater than a factor of 3. Because production of isoprene by tropical trees and lianas strongly increases at higher temperatures (Keller and Lerdau, 1999), tropical warming is likely to increase this NPP constituent.

Opportunities for data–model fusion will be maximized by developing the C cycle models to explicitly specify those NPP components that have been assessed in the field. As recently reported by Negrón-Juárez et al. (2015), only three of the 10 ESMs in CMIP5 report “leaf NPP”, “wood NPP”, and “root NPP”. The different production components are functionally distinct. In a landscape-scale field study at the Costa Rican LS site, the several field-quantified NPP components varied independently through 12 years, showing distinct relationships to the interannual variation in temperature, rainfall, and VPD (Clark et al., 2013). Below, we individually consider those biometric NPP components that have been assessed to date in tropical lowland forests.

In tropical forests, biometric aboveground NPP is typically dominated by short-lived tissues (Clark et al., 2001b). These are assayed as shed “fine litterfall” collected in litter traps (Table 8). Fine litterfall varies spatially within each tropical forest. When assessed in 18 0.5 ha plots distributed within one neotropical forest (LS, Table 8), the plots differed (max - min) in annual leaf litterfall by 3.8 to 6.3 Mgha-1yr-1, depending on the year; for reproductive litterfall, the across-plot range was >2 Mgha-1yr-1 in most of the 12 years (data in Table S2 in Clark et al., 2013). Landscape-scale data are therefore needed for reference-level benchmarks for this aspect of tropical-forest C cycling. Because the three components of fine litterfall are functionally distinct, they are considered individually below.

In field studies of biometric NPP (termed NPP*; Clark et al., 2001a), leaf litterfall over a given study interval is typically taken as a surrogate for leaf production over that interval. Stand-level leaf production itself has not been quantified in the field in tropical forests. In most tropical forests, leaf litterfall is the largest contributor to aboveground NPP* (Clark et al., 2013, and included references). It can be a misleading surrogate for leaf production in terms of both mass and timing. One method issue is the difficulty of quantifying the very large fallen leaves in tropical forests (e.g., 3 m long palm leaves). Ground-level and/or very large traps are required to collect these large items of fine litter (Villela and Proctor, 1999) but are rarely used. In addition, in tropical forests leaf litterfall undervalues leaf production due to two types of pre-collection losses (Table 7; also see Clark et al., 2001b). One is the mass loss from pre-collection decomposition and leaching of the shed leaves in the hot, humid conditions. Some leaves hang up in the vegetation and decompose above the ground. When Frangi and Lugo (1985) suspended old leaves from palms in a Puerto Rican forest, they found that roughly half the leaf mass was lost through decomposition in 4 months. A second issue is the leaf mass removed by herbivores (Table 7). Partial leaf damage (holes in fallen leaves) was estimated at ca. 0.8 Mg C ha-1yr-1 in a lowland Peruvian forest (Metcalfe et al., 2013); in addition, leaf-monitoring studies (Lowman, 1984; Filip et al., 1995) have shown that an equivalent amount or more may typically be lost to herbivores that remove entire leaves.

One potential approach for models would be to explicitly include the processes of herbivory and decomposition losses that occur between leaf production and leaf shedding, therefore facilitating a direct comparison. In lieu of this, model–data comparisons should take into account the low bias of leaf litterfall observations. In cases in which leaf litterfall is conflated with leaf production for the purposes of determining allocation to the leaf fraction, the resulting allocation underestimate might lead to underestimation of LAI.

A separate issue is that the seasonal timing of leaf production can differ from that of leaf litterfall, as found by Reich et al. (2004) in a Venezuelan tropical forest (in most species studied, although there was some degree of correlation). In many tropical forests, leaf litterfall typically peaks at the time of the yearly maximum soil dry-down (Wagner et al., 2016); this timing can be distinct from that of actual leaf production. Such a timing disjunct will complicate attempts to evaluate the seasonality of tropical-forest NPP and C allocation when leaf litterfall is used as the surrogate for production (e.g., Doughty et al., 2014).

Estimates of twig litterfall should be treated as a lower bound for twig production. In tropical forests, twig litterfall (Table 8) is likely to strongly underestimate actual production due to substantial mass loss before collection. In a New Guinea rain forest, when Edwards (1977) compared canopy-collected live twigs with a diameter <1 to <1 cm diameter twigs in the litter traps, the fallen twigs were found to have already lost 36–40 % of their mass, presumably due to decomposition and/or leaching when they were still attached to the branches above.

The biometric surrogate for reproductive production, reproductive litterfall (Table 8), is likely to undervalue production by at least 50 %. This NPP component is not easily quantified at the stand level. Tropical forests are typically dominated by animal-dispersed plants. The consumers are likely to remove most of the fruits produced, leaving the crumbs to fall into the litter traps. In a Puerto Rican palm forest, for example, fruit production assessed by direct observation over time exceeded the fruit mass in litter traps by a factor of 14 (Lugo and Frangi, 1993). Similarly, in a Colombian tropical forest, the estimate of fruit production based on observing from platforms and from climbing ropes was double the estimate based on fruit mass in the litter traps (Parrado-Rosselli et al., 2006).

For multiple reasons, this NPP component merits attention for the models. Many land surface models do not specifically include the carbon allocation to reproduction; this omission implies corresponding overestimates of stocks of other carbon pools (e.g., roots, stems, leaves). Demographic models, in contrast, typically do specify reproductive allocation, which is needed to drive forest recruitment (Moorcroft et al., 2001). Secondly, reproductive tissues are nutrient-rich (e.g., in nitrogen, phosphorus, and cations) and thus likely play a significant role in the cycling of those nutrients. Reproductive status could influence nutrient resorption and thus reallocation of carbon (Tully et al., 2013). A third issue is that this production component could be responding to climatic and/or [CO2] changes. Two recent tropical-forest studies suggest multi-decadal increases in forest-level reproduction (reproductive litterfall, Clark et al., 2013; flowering incidence, Pau et al., 2013).

As for aboveground woody biomass (above), field estimates of aboveground wood production, also termed EABI (estimated aboveground biomass increment), are unverified and highly uncertain. This production component is based on measurements at two successive censuses of the diameters of all live stems in the study plot that exceed an arbitrary diameter limit (usually 10 cm); these data are then used for allometric estimation of the tree's aboveground biomass at both times. EABI is calculated as the sum of the estimated biomass increments of all the stems that survived the interval, plus the estimated increments above the specified size limit of the recruits, those smaller stems that grew past the minimum size by the second census (see Clark et al., 2001a). One method variant (Chave et al., 2008b; Pyle et al., 2008), equating the census-interval growth of new recruits to their total estimated mass at the second census, substantially overestimates these small trees' contribution to stand growth; before reaching the 10 cm diameter limit, most small trees in tropical forests have grown very slowly over decades (see Clark and Clark, 2001; Rozendaal et al., 2015).

As for estimates of aboveground biomass, because EABI depends on an unverified allometric relationship between stem diameter and stem biomass, all values of this metric involve unquantifiable uncertainty. When different allometries are applied to the same set of diameter data, different estimates of EABI can be produced (e.g., duplicate estimates at site TAP-KM67; Table 9). Determining which if any of such estimates are reasonable would require follow-up on-site verification of the underlying allometry (Clark and Kellner, 2012).

Given the heterogeneity of biomass dynamics within a tropical forest, data–model fusion exercises and site-level model testing call for landscape-scale field data for EABI. Individual-based or demographic models (e.g., ED, Moorcroft et al., 2001) that address the small-scale spatial heterogeneity within a forest landscape are the exceptions to this. In spite of this metric's unquantifiable uncertainty, when estimated on the landscape scale and in the same way over a long series of successive periods, repeated annual estimates can provide valuable guidance for the models with respect to both long-term trends in this productivity component and its climatic and [CO2] responses. For example, 12-year records of EABI from the LS site revealed highly significant sensitivities of landscape-scale EABI to the inter-year changes in nighttime temperatures, VPD, and [CO2] (Clark et al., 2013).

Field estimates of fine-root production at the landscape level in tropical forests provide a useful lower bound for this NPP component. Due to the method challenges, fine-root production has not been well quantified in any forest type, boreal to tropical. In the tropical-forest biome, because of the notorious variation in fine-root stocks on all spatial scales (Espeleta and Clark, 2007; Powers et al., 2005), robust assessment of fine-root production for a given forest would require highly replicated and distributed sampling. Unfortunately, this production component has only rarely been assessed in multiple hectares of a tropical forest (Table 10). A second critical limitation is that the field measurements to date in this biome have been confined to the surface soil (0 to ≤30 cm depth). There are no field observations from tropical forests of production by the deeper fine roots (live fine roots were found to at least 18 m depth in one Amazon forest; Nepstad et al., 1994).

Variable methods for assessing fine-root production (different soil depths and root sizes, inclusion or exclusion of dead roots; Table 10) also make cross-site comparisons difficult. The usual approach in tropical forests, in-growth cores, is likely to strongly underestimate production due to lags before root in-growth and the likelihood of roots dying and decomposing before soil cores are retrieved; in a temperate pine forest, production estimates based on in-growth cores averaged 54 % lower than those from minirhizotrons (Hendricks et al., 2006). Whether root herbivory removes a significant fraction of fine-root production (Lauenroth, 2000) is as yet unstudied in tropical forests.